This research investigates how students with varying academic achievement levels utilize ChatGPT in the process of writing their undergraduate theses. Employing a qualitative phenomenological approach, the study explores the patterns of use, perceptions, motivations, and challenges experienced by high-achieving, mid-achieving, and low-achieving students. Data were collected through in-depth interviews with three participants from Universitas Islam Malang, each representing a distinct academic category. The findings reveal that high-achieving students use ChatGPT critically and selectively to refine coherence and clarity, while mid-achievers rely on it for structural organization and motivational support. In contrast, low-achieving students exhibit a higher degree of dependence, often using ChatGPT as a primary source of content and guidance. Across all groups, ChatGPT is perceived as both a cognitive and emotional aid, though ethical concerns such as originality, over-reliance, and learning depth vary significantly. The research highlights the need for differentiated pedagogical strategies and ethical guidance to ensure responsible AI integration in academic writing.
Copyrights © 2025